PIX *
MakeReplacementMask(PIX *pixs)
{
PIX *pix1, *pix2, *pix3, *pix4;
pix1 = pixMaskOverColorPixels(pixs, 95, 3);
pix2 = pixMorphSequence(pix1, "o15.15", 0);
pixSeedfillBinary(pix2, pix2, pix1, 8);
pix3 = pixMorphSequence(pix2, "c15.15 + d61.31", 0);
pix4 = pixRemoveBorderConnComps(pix3, 8);
pixXor(pix4, pix4, pix3);
pixDestroy(&pix1);
pixDestroy(&pix2);
pixDestroy(&pix3);
return pix4;
}
/*!
* pixThinExamples()
*
* Input: pixs (1 bpp)
* type (L_THIN_FG, L_THIN_BG)
* index (into specific examples; valid 1-9; see notes)
* maxiters (max number of iters allowed; use 0 to iterate
* until completion)
* selfile (<optional> filename for output sel display)
* Return: pixd, or null on error
*
* Notes:
* (1) See notes in pixThin(). The examples are taken from
* the paper referenced there.
* (2) Here we allow specific sets of HMTs to be used in
* parallel for thinning from each of four directions.
* One iteration consists of four such parallel thins.
* (3) The examples are indexed as follows:
* Thinning (e.g., run to completion):
* index = 1 sel_4_1, sel_4_5, sel_4_6
* index = 2 sel_4_1, sel_4_7, sel_4_7_rot
* index = 3 sel_48_1, sel_48_1_rot, sel_48_2
* index = 4 sel_8_2, sel_8_3, sel_48_2
* index = 5 sel_8_1, sel_8_5, sel_8_6
* index = 6 sel_8_2, sel_8_3, sel_8_8, sel_8_9
* index = 7 sel_8_5, sel_8_6, sel_8_7, sel_8_7_rot
* Thickening:
* index = 8 sel_4_2, sel_4_3 (e.g,, do just a few iterations)
* index = 9 sel_8_4 (e.g., do just a few iterations)
*/
PIX *
pixThinExamples(PIX *pixs,
l_int32 type,
l_int32 index,
l_int32 maxiters,
const char *selfile)
{
PIX *pixd, *pixt;
SEL *sel;
SELA *sela;
PROCNAME("pixThinExamples");
if (!pixs)
return (PIX *)ERROR_PTR("pixs not defined", procName, NULL);
if (pixGetDepth(pixs) != 1)
return (PIX *)ERROR_PTR("pixs not 1 bpp", procName, NULL);
if (type != L_THIN_FG && type != L_THIN_BG)
return (PIX *)ERROR_PTR("invalid fg/bg type", procName, NULL);
if (index < 1 || index > 9)
return (PIX *)ERROR_PTR("invalid index", procName, NULL);
if (maxiters == 0) maxiters = 10000;
switch(index)
{
case 1:
sela = selaCreate(3);
sel = selCreateFromString(sel_4_1, 3, 3, "sel_4_1");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_4_5, 3, 3, "sel_4_5");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_4_6, 3, 3, "sel_4_6");
selaAddSel(sela, sel, NULL, 0);
break;
case 2:
sela = selaCreate(3);
sel = selCreateFromString(sel_4_1, 3, 3, "sel_4_1");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_4_7, 3, 3, "sel_4_7");
selaAddSel(sela, sel, NULL, 0);
sel = selRotateOrth(sel, 1);
selaAddSel(sela, sel, "sel_4_7_rot", 0);
break;
case 3:
sela = selaCreate(3);
sel = selCreateFromString(sel_48_1, 3, 3, "sel_48_1");
selaAddSel(sela, sel, NULL, 0);
sel = selRotateOrth(sel, 1);
selaAddSel(sela, sel, "sel_48_1_rot", 0);
sel = selCreateFromString(sel_48_2, 3, 3, "sel_48_2");
selaAddSel(sela, sel, NULL, 0);
break;
case 4:
sela = selaCreate(3);
sel = selCreateFromString(sel_8_2, 3, 3, "sel_8_2");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_3, 3, 3, "sel_8_3");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_48_2, 3, 3, "sel_48_2");
selaAddSel(sela, sel, NULL, 0);
break;
case 5:
sela = selaCreate(3);
sel = selCreateFromString(sel_8_1, 3, 3, "sel_8_1");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_5, 3, 3, "sel_8_5");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_6, 3, 3, "sel_8_6");
selaAddSel(sela, sel, NULL, 0);
break;
case 6:
sela = selaCreate(4);
sel = selCreateFromString(sel_8_2, 3, 3, "sel_8_2");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_3, 3, 3, "sel_8_3");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_8, 3, 3, "sel_8_8");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_9, 3, 3, "sel_8_9");
selaAddSel(sela, sel, NULL, 0);
break;
case 7:
sela = selaCreate(4);
sel = selCreateFromString(sel_8_5, 3, 3, "sel_8_5");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_6, 3, 3, "sel_8_6");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_8_7, 3, 3, "sel_8_7");
selaAddSel(sela, sel, NULL, 0);
sel = selRotateOrth(sel, 1);
selaAddSel(sela, sel, "sel_8_7_rot", 0);
break;
case 8: /* thicken for this one; just a few iterations */
sela = selaCreate(2);
sel = selCreateFromString(sel_4_2, 3, 3, "sel_4_2");
selaAddSel(sela, sel, NULL, 0);
sel = selCreateFromString(sel_4_3, 3, 3, "sel_4_3");
selaAddSel(sela, sel, NULL, 0);
pixt = pixThinGeneral(pixs, type, sela, maxiters);
pixd = pixRemoveBorderConnComps(pixt, 4);
pixDestroy(&pixt);
break;
case 9: /* thicken for this one; just a few iterations */
sela = selaCreate(1);
sel = selCreateFromString(sel_8_4, 3, 3, "sel_8_4");
selaAddSel(sela, sel, NULL, 0);
pixt = pixThinGeneral(pixs, type, sela, maxiters);
pixd = pixRemoveBorderConnComps(pixt, 4);
pixDestroy(&pixt);
break;
default:
return (PIX *)ERROR_PTR("invalid index", procName, NULL);
}
if (index <= 7)
pixd = pixThinGeneral(pixs, type, sela, maxiters);
/* Optionally display the sels */
if (selfile) {
pixt = selaDisplayInPix(sela, 35, 3, 15, 4);
pixWrite(selfile, pixt, IFF_PNG);
pixDestroy(&pixt);
}
selaDestroy(&sela);
return pixd;
}
/*!
* pixFindPageForeground()
*
* Input: pixs (full resolution (any type or depth)
* threshold (for binarization; typically about 128)
* mindist (min distance of text from border to allow
* cleaning near border; at 2x reduction, this
* should be larger than 50; typically about 70)
* erasedist (when conditions are satisfied, erase anything
* within this distance of the edge;
* typically 30 at 2x reduction)
* pagenum (use for debugging when called repeatedly; labels
* debug images that are assembled into pdfdir)
* showmorph (set to a negative integer to show steps in
* generating masks; this is typically used
* for debugging region extraction)
* display (set to 1 to display mask and selected region
* for debugging a single page)
* pdfdir (subdirectory of /tmp where images showing the
* result are placed when called repeatedly; use
* null if no output requested)
* Return: box (region including foreground, with some pixel noise
* removed), or null if not found
*
* Notes:
* (1) This doesn't simply crop to the fg. It attempts to remove
* pixel noise and junk at the edge of the image before cropping.
* The input @threshold is used if pixs is not 1 bpp.
* (2) There are several debugging options, determined by the
* last 4 arguments.
* (3) If you want pdf output of results when called repeatedly,
* the pagenum arg labels the images written, which go into
* /tmp/<pdfdir>/<pagenum>.png. In that case,
* you would clean out the /tmp directory before calling this
* function on each page:
* lept_rmdir(pdfdir);
* lept_mkdir(pdfdir);
*/
BOX *
pixFindPageForeground(PIX *pixs,
l_int32 threshold,
l_int32 mindist,
l_int32 erasedist,
l_int32 pagenum,
l_int32 showmorph,
l_int32 display,
const char *pdfdir)
{
char buf[64];
l_int32 flag, nbox, intersects;
l_int32 w, h, bx, by, bw, bh, left, right, top, bottom;
PIX *pixb, *pixb2, *pixseed, *pixsf, *pixm, *pix1, *pixg2;
BOX *box, *boxfg, *boxin, *boxd;
BOXA *ba1, *ba2;
PROCNAME("pixFindPageForeground");
if (!pixs)
return (BOX *)ERROR_PTR("pixs not defined", procName, NULL);
/* Binarize, downscale by 0.5, remove the noise to generate a seed,
* and do a seedfill back from the seed into those 8-connected
* components of the binarized image for which there was at least
* one seed pixel. Also clear out any components that are within
* 10 pixels of the edge at 2x reduction. */
flag = (showmorph) ? -1 : 0; /* if showmorph == -1, write intermediate
* images to /tmp/seq_output_1.pdf */
pixb = pixConvertTo1(pixs, threshold);
pixb2 = pixScale(pixb, 0.5, 0.5);
pixseed = pixMorphSequence(pixb2, "o1.2 + c9.9 + o3.5", flag);
pixsf = pixSeedfillBinary(NULL, pixseed, pixb2, 8);
pixSetOrClearBorder(pixsf, 10, 10, 10, 10, PIX_SET);
pixm = pixRemoveBorderConnComps(pixsf, 8);
if (display) pixDisplay(pixm, 100, 100);
/* Now, where is the main block of text? We want to remove noise near
* the edge of the image, but to do that, we have to be convinced that
* (1) there is noise and (2) it is far enough from the text block
* and close enough to the edge. For each edge, if the block
* is more than mindist from that edge, then clean 'erasedist'
* pixels from the edge. */
pix1 = pixMorphSequence(pixm, "c50.50", flag - 1);
ba1 = pixConnComp(pix1, NULL, 8);
ba2 = boxaSort(ba1, L_SORT_BY_AREA, L_SORT_DECREASING, NULL);
pixGetDimensions(pix1, &w, &h, NULL);
nbox = boxaGetCount(ba2);
if (nbox > 1) {
box = boxaGetBox(ba2, 0, L_CLONE);
boxGetGeometry(box, &bx, &by, &bw, &bh);
left = (bx > mindist) ? erasedist : 0;
right = (w - bx - bw > mindist) ? erasedist : 0;
top = (by > mindist) ? erasedist : 0;
bottom = (h - by - bh > mindist) ? erasedist : 0;
pixSetOrClearBorder(pixm, left, right, top, bottom, PIX_CLR);
boxDestroy(&box);
}
pixDestroy(&pix1);
boxaDestroy(&ba1);
boxaDestroy(&ba2);
/* Locate the foreground region; don't bother cropping */
pixClipToForeground(pixm, NULL, &boxfg);
/* Sanity check the fg region. Make sure it's not confined
* to a thin boundary on the left and right sides of the image,
* in which case it is likely to be noise. */
if (boxfg) {
boxin = boxCreate(0.1 * w, 0, 0.8 * w, h);
boxIntersects(boxfg, boxin, &intersects);
if (!intersects) {
L_INFO("found only noise on page %d\n", procName, pagenum);
boxDestroy(&boxfg);
}
boxDestroy(&boxin);
}
boxd = NULL;
if (!boxfg) {
L_INFO("no fg region found for page %d\n", procName, pagenum);
} else {
boxAdjustSides(boxfg, boxfg, -2, 2, -2, 2); /* tiny expansion */
boxd = boxTransform(boxfg, 0, 0, 2.0, 2.0);
/* Write image showing box for this page. This is to be
* bundled up into a pdf of all the pages, which can be
* generated by convertFilesToPdf() */
if (pdfdir) {
pixg2 = pixConvert1To4Cmap(pixb);
pixRenderBoxArb(pixg2, boxd, 3, 255, 0, 0);
snprintf(buf, sizeof(buf), "/tmp/%s/%05d.png", pdfdir, pagenum);
if (display) pixDisplay(pixg2, 700, 100);
pixWrite(buf, pixg2, IFF_PNG);
pixDestroy(&pixg2);
}
}
pixDestroy(&pixb);
pixDestroy(&pixb2);
pixDestroy(&pixseed);
pixDestroy(&pixsf);
pixDestroy(&pixm);
boxDestroy(&boxfg);
return boxd;
}
/*!
* Note: this method is generally inferior to pixHasColorRegions(); it
* is retained as a reference only
*
* \brief pixFindColorRegionsLight()
*
* \param[in] pixs 32 bpp rgb
* \param[in] pixm [optional] 1 bpp mask image
* \param[in] factor subsample factor; integer >= 1
* \param[in] darkthresh threshold to eliminate dark pixels (e.g., text)
* from consideration; typ. 70; -1 for default.
* \param[in] lightthresh threshold for minimum gray value at 95% rank
* near white; typ. 220; -1 for default
* \param[in] mindiff minimum difference from 95% rank value, used
* to count darker pixels; typ. 50; -1 for default
* \param[in] colordiff minimum difference in (max - min) component to
* qualify as a color pixel; typ. 40; -1 for default
* \param[out] pcolorfract fraction of 'color' pixels found
* \param[out] pcolormask1 [optional] mask over background color, if any
* \param[out] pcolormask2 [optional] filtered mask over background color
* \param[out] pixadb [optional] debug intermediate results
* \return 0 if OK, 1 on error
*
* <pre>
* Notes:
* (1) This function tries to determine if there is a significant
* color or darker region on a scanned page image where part
* of the image is very close to "white". It will also allow
* extraction of small regions of lightly colored pixels.
* If the background is darker (and reddish), use instead
* pixHasColorRegions2().
* (2) If %pixm exists, only pixels under fg are considered. Typically,
* the inverse of %pixm would have fg pixels over a photograph.
* (3) There are four thresholds.
* * %darkthresh: ignore pixels darker than this (typ. fg text).
* We make a 1 bpp mask of these pixels, and then dilate it to
* remove all vestiges of fg from their vicinity.
* * %lightthresh: let val95 be the pixel value for which 95%
* of the non-masked pixels have a lower value (darker) of
* their min component. Then if val95 is darker than
* %lightthresh, the image is not considered to have a
* light bg, and this returns 0.0 for %colorfract.
* * %mindiff: we are interested in the fraction of pixels that
* have two conditions. The first is that their min component
* is at least %mindiff darker than val95.
* * %colordiff: the second condition is that the max-min diff
* of the pixel components exceeds %colordiff.
* (4) This returns in %pcolorfract the fraction of pixels that have
* both a min component that is at least %mindiff below that at the
* 95% rank value (where 100% rank is the lightest value), and
* a max-min diff that is at least %colordiff. Without the
* %colordiff constraint, gray pixels of intermediate value
* could get flagged by this function.
* (5) No masks are returned unless light color pixels are found.
* If colorfract > 0.0 and %pcolormask1 is defined, this returns
* a 1 bpp mask with fg pixels over the color background.
* This mask may have some holes in it.
* (6) If colorfract > 0.0 and %pcolormask2 is defined, this returns
* a filtered version of colormask1. The two changes are
* (a) small holes have been filled
* (b) components near the border have been removed.
* The latter insures that dark pixels near the edge of the
* image are not included.
* (7) To generate a boxa of rectangular regions from the overlap
* of components in the filtered mask:
* boxa1 = pixConnCompBB(colormask2, 8);
* boxa2 = boxaCombineOverlaps(boxa1);
* This is done here in debug mode.
* </pre>
*/
static l_int32
pixFindColorRegionsLight(PIX *pixs,
PIX *pixm,
l_int32 factor,
l_int32 darkthresh,
l_int32 lightthresh,
l_int32 mindiff,
l_int32 colordiff,
l_float32 *pcolorfract,
PIX **pcolormask1,
PIX **pcolormask2,
PIXA *pixadb)
{
l_int32 lightbg, w, h, count;
l_float32 ratio, val95, rank;
BOXA *boxa1, *boxa2;
NUMA *nah;
PIX *pix1, *pix2, *pix3, *pix4, *pix5, *pixm1, *pixm2, *pixm3;
PROCNAME("pixFindColorRegionsLight");
if (pcolormask1) *pcolormask1 = NULL;
if (pcolormask2) *pcolormask2 = NULL;
if (!pcolorfract)
return ERROR_INT("&colorfract not defined", procName, 1);
*pcolorfract = 0.0;
if (!pixs || pixGetDepth(pixs) != 32)
return ERROR_INT("pixs not defined or not 32 bpp", procName, 1);
if (factor < 1) factor = 1;
if (darkthresh < 0) darkthresh = 70; /* defaults */
if (lightthresh < 0) lightthresh = 220;
if (mindiff < 0) mindiff = 50;
if (colordiff < 0) colordiff = 40;
/* Check if pixm covers most of the image. If so, just return. */
pixGetDimensions(pixs, &w, &h, NULL);
if (pixm) {
pixCountPixels(pixm, &count, NULL);
ratio = (l_float32)count / ((l_float32)(w) * h);
if (ratio > 0.7) {
if (pixadb) L_INFO("pixm has big fg: %f5.2\n", procName, ratio);
return 0;
}
}
/* Make a mask pixm1 over the dark pixels in the image:
* convert to gray using the average of the components;
* threshold using %darkthresh; do a small dilation;
* combine with pixm. */
pix1 = pixConvertRGBToGray(pixs, 0.33, 0.34, 0.33);
if (pixadb) pixaAddPix(pixadb, pixs, L_COPY);
if (pixadb) pixaAddPix(pixadb, pix1, L_COPY);
pixm1 = pixThresholdToBinary(pix1, darkthresh);
pixDilateBrick(pixm1, pixm1, 7, 7);
if (pixadb) pixaAddPix(pixadb, pixm1, L_COPY);
if (pixm) {
pixOr(pixm1, pixm1, pixm);
if (pixadb) pixaAddPix(pixadb, pixm1, L_COPY);
}
pixDestroy(&pix1);
/* Convert to gray using the minimum component value and
* find the gray value at rank 0.95, that represents the light
* pixels in the image. If it is too dark, quit. */
pix1 = pixConvertRGBToGrayMinMax(pixs, L_SELECT_MIN);
pix2 = pixInvert(NULL, pixm1); /* pixels that are not dark */
pixGetRankValueMasked(pix1, pix2, 0, 0, factor, 0.95, &val95, &nah);
pixDestroy(&pix2);
if (pixadb) {
L_INFO("val at 0.95 rank = %5.1f\n", procName, val95);
gplotSimple1(nah, GPLOT_PNG, "/tmp/lept/histo1", "gray histo");
pix3 = pixRead("/tmp/lept/histo1.png");
pix4 = pixExpandReplicate(pix3, 2);
pixaAddPix(pixadb, pix4, L_INSERT);
pixDestroy(&pix3);
}
lightbg = (l_int32)val95 >= lightthresh;
numaDestroy(&nah);
if (!lightbg) {
pixDestroy(&pix1);
pixDestroy(&pixm1);
return 0;
}
/* Make mask pixm2 over pixels that are darker than val95 - mindiff. */
pixm2 = pixThresholdToBinary(pix1, val95 - mindiff);
if (pixadb) pixaAddPix(pixadb, pixm2, L_COPY);
pixDestroy(&pix1);
/* Make a mask pixm3 over pixels that have some color saturation,
* with a (max - min) component difference >= %colordiff,
* and combine using AND with pixm2. */
pix2 = pixConvertRGBToGrayMinMax(pixs, L_CHOOSE_MAXDIFF);
pixm3 = pixThresholdToBinary(pix2, colordiff);
pixDestroy(&pix2);
pixInvert(pixm3, pixm3); /* need pixels above threshold */
if (pixadb) pixaAddPix(pixadb, pixm3, L_COPY);
pixAnd(pixm2, pixm2, pixm3);
if (pixadb) pixaAddPix(pixadb, pixm2, L_COPY);
pixDestroy(&pixm3);
/* Subtract the dark pixels represented by pixm1.
* pixm2 now holds all the color pixels of interest */
pixSubtract(pixm2, pixm2, pixm1);
pixDestroy(&pixm1);
if (pixadb) pixaAddPix(pixadb, pixm2, L_COPY);
/* But we're not quite finished. Remove pixels from any component
* that is touching the image border. False color pixels can
* sometimes be found there if the image is much darker near
* the border, due to oxidation or reduced illumination. */
pixm3 = pixRemoveBorderConnComps(pixm2, 8);
pixDestroy(&pixm2);
if (pixadb) pixaAddPix(pixadb, pixm3, L_COPY);
/* Get the fraction of light color pixels */
pixCountPixels(pixm3, &count, NULL);
*pcolorfract = (l_float32)count / (w * h);
if (pixadb) {
if (count == 0)
L_INFO("no light color pixels found\n", procName);
else
L_INFO("fraction of light color pixels = %5.3f\n", procName,
*pcolorfract);
}
/* Debug: extract the color pixels from pixs */
if (pixadb && count > 0) {
/* Use pixm3 to extract the color pixels */
pix3 = pixCreateTemplate(pixs);
pixSetAll(pix3);
pixCombineMasked(pix3, pixs, pixm3);
pixaAddPix(pixadb, pix3, L_INSERT);
/* Use additional filtering to extract the color pixels */
pix3 = pixCloseSafeBrick(NULL, pixm3, 15, 15);
pixaAddPix(pixadb, pix3, L_INSERT);
pix5 = pixCreateTemplate(pixs);
pixSetAll(pix5);
pixCombineMasked(pix5, pixs, pix3);
pixaAddPix(pixadb, pix5, L_INSERT);
/* Get the combined bounding boxes of the mask components
* in pix3, and extract those pixels from pixs. */
boxa1 = pixConnCompBB(pix3, 8);
boxa2 = boxaCombineOverlaps(boxa1, NULL);
pix4 = pixCreateTemplate(pix3);
pixMaskBoxa(pix4, pix4, boxa2, L_SET_PIXELS);
pixaAddPix(pixadb, pix4, L_INSERT);
pix5 = pixCreateTemplate(pixs);
pixSetAll(pix5);
pixCombineMasked(pix5, pixs, pix4);
pixaAddPix(pixadb, pix5, L_INSERT);
boxaDestroy(&boxa1);
boxaDestroy(&boxa2);
pixaAddPix(pixadb, pixs, L_COPY);
}
/* Optional colormask returns */
if (pcolormask2 && count > 0)
*pcolormask2 = pixCloseSafeBrick(NULL, pixm3, 15, 15);
if (pcolormask1 && count > 0)
*pcolormask1 = pixm3;
else
pixDestroy(&pixm3);
return 0;
}